For years, tech companies thrived by scaling software at near-zero cost.
AI is changing that: it’s easier to build products, but more expensive to run them. That shift could reshape profits, competition and what investors value most.
Over the past decade, investment markets have richly rewarded companies built on intangible capital – non-physical assets that contribute to the value, productivity, or earning potential of a business – particularly those leveraging intellectual property and network effects. This has been most evident in software, data, brands and platforms. Once established, these businesses can scale to millions of users without requiring significant additional investment. The result has been the rise of a small number of very large technology firms generating persistent, highly scalable profits. Stocks in this category include the likes of Alphabet, Meta, Microsoft and Nvidia.
What is often missed is how the rise of these businesses dovetailed perfectly with the broader macroeconomic environment. In the aftermath of the 2008 financial crisis, central banks aggressively lowered interest rates, while quantitative easing helped anchor them at low levels. This compressed discount rates – the rate that central banks charge other financial institutions to borrow – making long-duration profit streams particularly valuable. Money flowing into passive funds strengthened this effect, as market-cap weighting directs more money to the largest listed firms. As valuations rose, their cost of capital fell, enabling further investment and expansion. In this way, past success fed into future dominance, creating a reinforcing loop between monetary policy, index construction and the economics of intangible assets.
Advantage to the giants
At the centre of the model was a distinctive cost structure: high fixed costs and very low variable costs, which encouraged firms to prioritise rapid growth and market share. In other words, because adding new customers was seen as relatively cheap, these companies could prioritise growing as fast as possible, even if they lost money at first. A moto popularised by Meta’s Mark Zuckerberg was to “move fast and break things” – encouraging engineers to prioritise speed, risk-taking, and innovation over caution or perfection.
Firms that this adopted approach were able to reach critical mass, which in turn helped sustain margins and defend market share over long periods. Low interest rates amplified these advantages further, as the value of distant future earnings rose when discount rates fell.
Large technology platforms may also have contributed to a disinflationary backdrop by putting downward pressure on prices. Their advertising-funded services undercut or eliminated traditional pricing (for example, Google Maps vs. paid navigation, YouTube vs. cable TV) and marketplaces like Amazon reduced information asymmetries and forced price competition across retailers. That, in turn, gave central banks scope to keep policy accommodative for longer, reinforcing the same conditions that supported these firms’ growth. The interaction between near-zero marginal costs, passive capital allocation and compressed interest rates created a self-reinforcing regime.
Cheaper to build, expensive to serve
AI does something unusual: it pushes the economics in two directions at once, and the net effect on profitability is far less straightforward than either the optimists or the sceptics tend to acknowledge.
On one side, AI lowers the cost of building software. Code can be written faster, products can be launched with fewer people, and functionality can be added more cheaply. The upfront investment required to create a viable product falls. That weakens one of the old era’s most powerful defences: if building is cheaper, barriers to entry are lower, and incumbents face more frequent competitive challenge.
On the other side, AI raises the cost of serving customers. Running machine learning at scale is expensive. It depends on computing power and energy, semiconductors, cooling systems and data-centre capacity. Each additional unit of usage carries real incremental cost in a way that a traditional software product did not.
The economic model that results from these two forces is quite different from the one that investors grew comfortable with over the past decade. Previously, this depended on high fixed costs spread across a growing user base at near-zero marginal cost. AI inverts this: lower build costs but serving costs that rise as usage grows. We believe investors have been slow to internalise the full implications: AI companies may never enjoy the effortless margin expansion that defined the last technology cycle.
AI makes digital products physical again
AI products cannot scale without a concomitant scaling of the physical infrastructure: chips, data centres, grid connections, power generation and cooling infrastructure. Previously, this meant value accrued to platforms and software companies that owned hard-to-replicate intangible assets. In the years ahead, it may shift towards those who can finance, build, operate, and control physical infrastructure at scale.
Software, data and network effects have not ceased to matter. The change is that returns are no longer purely asset-light. Profitability increasingly depends on a combination of intangible capability and tangible capacity – in other words, a business needs both ‘soft’ strengths and ‘hard’ assets working together. Firms that lack access to the tangible side may find their economics deteriorating even as demand for their products grows.
AI makes competition easier, but scale more expensive
Because AI lowers the cost of creating software, entry barriers fall. Because serving costs rise with usage, scale becomes less automatically profitable. A company could attract strong demand yet still fail to generate the margin expansion that investors learned to expect during the platform era. Heavy adoption may be unattractive if each additional user imposes real incremental cost on the business. Acquisitions of upstart competitors, a familiar tool for defending margins, may also become less effective. If AI keeps the cost of building new products low, acquired competitors can be replaced more easily by new entrants. The persistence of profits weakens on both sides: margins compress because serving costs are higher, and competitive moats narrow because building costs are lower. This creates serious valuation risk. Investors applying assumptions formed during the intangible era – high margins, durable moats, minimal ongoing capex – to AI businesses will systematically overestimate returns and underestimate capital requirements.
Reinforcing the shift
The macroeconomic environment reinforces this structural shift. Since the pandemic, the balance between monetary and fiscal policy has moved, and the world of zero interest rates and repeated quantitative easing appears unlikely to return. Fiscal pressures have risen across developed economies, inflation risk looks more persistent, and the hurdle for renewed monetary accommodation is higher than it was in the decade after the financial crisis.
In that environment, the valuation support that long-duration intangible profit streams received during the previous decade may be less reliable. Compressed discount rates made those streams exceptionally valuable. If rates remain structurally higher, the automatic uplift fades.
At the same time, a world shaped more by industrial policy, infrastructure investment, and strategic competition between major economies may prove more supportive of businesses linked to physical capacity and supply-side investment. The shift from intangible profits toward tangible capital is occurring precisely as the macro environment shifts to favour such assets.
What does this mean for investors?
During the intangible era, the central investment question was: which firms own the most valuable intangible assets? Going forward, the more productive question is: which firms can combine intangible capability with access to scarce tangible inputs? This includes inputs such as computing power, semiconductors, physical infrastructure and the financing capacity to deploy them at scale.
Some incumbents will emerge stronger from this transition because they can fund infrastructure and absorb the cost of scaling. Others may discover that AI makes their business models more capital-intensive and less profitable than their historical track record suggests. Parts of the supply chain once considered peripheral may capture a greater share of economic value.
The post-crisis period rewarded intangible capital, near-zero marginal cost and scalable profits. The AI era appears to depend more on tangible investment, physical bottlenecks and the ability to convert digital capability into economically viable production. Investors may be entering not just a new technology cycle, but a different capital environment altogether.
Important information
This document is intended for retail investors and/or private clients. You should not act or rely on this document but should contact your professional adviser.
This document has been issued by Sarasin & Partners LLP of 50 George Street, London W1U 7DY, a limited liability partnership registered in England and Wales with registered number OC329859, and which is authorised and regulated by the Financial Conduct Authority with firm reference number 475111.
This document has been prepared for marketing and information purposes only and is not a solicitation, or an offer to buy or sell any security. The information on which the material is based has been obtained in good faith, from sources that we believe to be reliable, but we have not independently verified such information and we make no representation or warranty, express or implied, as to its accuracy. All expressions of opinion are subject to change without notice.
This document should not be relied on for accounting, legal or tax advice, or investment recommendations. Reliance should not be placed on the views and information in this material when taking individual investment and/or strategic decisions.
The value of investments and any income derived from them can fall as well as rise and investors may not get back the amount originally invested. If investing in foreign currencies, the return in the investor’s reference currency may increase or decrease as a result of currency fluctuations. Past performance is not a reliable indicator of future results and may not be repeated. Forecasts are not a reliable indicator of future performance.
Neither Sarasin & Partners LLP nor any other member of the J. Safra Sarasin Holding Ltd group accepts any liability or responsibility whatsoever for any consequential loss of any kind arising out of the use of this document or any part of its contents. The use of this document should not be regarded as a substitute for the exercise by the recipient of their own judgement. Sarasin & Partners LLP and/or any person connected with it may act upon or make use of the material referred to herein and/or any of the information upon which it is based, prior to publication of this document.
Where the data in this document comes partially from third-party sources the accuracy, completeness or correctness of the information contained in this publication is not guaranteed, and third-party data is provided without any warranties of any kind. Sarasin & Partners LLP shall have no liability in connection with third-party data.
© 2026 Sarasin & Partners LLP – all rights reserved. This document can only be distributed or reproduced with permission from Sarasin & Partners LLP. Please contact [email protected].